| 摘要 |
PURPOSE:To reproduce desired speech data only by inputting a speech for retrieval and to eliminate the need for a keyboard and a display device by storing speech pattern for retrieval and speech data to be retrieved and comparing them with the pattern of a retrieval input speech. CONSTITUTION:An analog speech signal which is inputted through a microphone 1a is passed through an LPF 2 and A/D-converted 3. Its output is stored on a RAM 5, patterned 11, and stored in a speech pattern RAM 13 through an address control circuit 14. Then when retrieval is carried out, the output, the inversion of Q of an FF 28, is held at high level and a speech of a desired name is inputted through the microphone 1a; and then the signal from an AND circuit 32 is supplied to the circuit 11 and stored temporarily on a storage memory 15. Here, a switch S4 is opened to detect and send the rising output of an inverter 39 to the circuit 14 and pattern data read out the storage area of the RAM 13 is compared with the output of a memory 15 and decided 16. When they coincide with each other, a coincidence signal L is sent to the RAM 5 and a desired reproduced signal is outputted. |
